[Solid_software] solid work disk is full now

Zhiwen Zhao zwzhao at jlab.org
Tue May 3 16:26:33 EDT 2022


All

the problem is fixed temporally after CC delete the ~1TB snapshot
We should still keep work disk usage clean and tight to minimize similar 
problems in future.

Zhiwen

On 5/3/2022 1:15 PM, Zhiwen Zhao wrote:
> Ole
> 
> I think we can delete all snapshot now.
> Could you help ask CC?
> 
> Then at work the disk will be usable again and we can go on with some 
> cleanup
> 
> Zhiwen
> 
> On 5/3/2022 12:51 PM, Ole Hansen wrote:
>> One problem you'll possibly encounter is that deleting files doesn't 
>> free up space. This may happen because of the snapshot feature of the 
>> work disks ZFS file system. The space will eventually be freed up as 
>> old snapshots are deleted automatically, but this can take up to a 
>> week. To free up space quickly, I'll have to ask the computer center 
>> to delete some snapshots manually.
>>
>> You can see the snapshots in the hidden directory 
>> /work/halla/solid/.zfs/snapshot/.  They can be useful in case of an 
>> accidental deletion.
>>
>> Does anyone currently care about any of the snapshots? If not, we can 
>> just delete all of them right now.
>>
>> Ole
>>
>> On 3.5.22 at 12:36, Zhiwen Zhao wrote:
>>> Dear All
>>>
>>> solid work disk is 100% full now so any attempt to write will fail
>>> Please delete/move large files and consider what can be put on 
>>> volatile or tape
>>> It's also very slow and make it hard to do it clean up.
>>> This might be a chicken-egg problem.
>>>
>>> Zhiwen
>>> _______________________________________________
>>> Solid_software mailing list
>>> Solid_software at jlab.org
>>> https://mailman.jlab.org/mailman/listinfo/solid_software
>>
>>
>> _______________________________________________
>> Solid_software mailing list
>> Solid_software at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/solid_software
> _______________________________________________
> Solid_software mailing list
> Solid_software at jlab.org
> https://mailman.jlab.org/mailman/listinfo/solid_software


More information about the Solid_software mailing list